Mastodawn

Mark Reinhold Mar 11

One thing I love about being involved in my College is I get a direct line on what young people are thinking.

In my circle of 30-50-somethings, it's taken for granted that "AI" is changing everything. Some people express caution that we shouldn't even take any position on software engineering education because it will be out of date within a week.

Meanwhile, my students are telling me that in their cohort, everyone is tired of AI and wants to learn proper software engineering.

One lesson here is: do you remember how clueless your boss was about everything? Recognise when you are in the "boss" demographic, and that your breathless enthusiasm and concern may be equally misinformed.

Jon Sterling Jan 26

My point of view is that generative AI is an interesting and powerful technology that has, due to the need to get a return on investment sooner rather than later, been deployed in harmful and counterproductive ways.

I cannot deny the strength of the technology, but I also cannot accept the breathless pronouncements that all seem to rely on assumptions that are emanating from a managerial stratum that has a lot of money and prestige riding on generative AI "breaking" all systems of production.

I find what some people are able to do using generative AI very interesting and inspiring. For example, @mitchellh is building Ghostty (a terminal application) to an extremely high standard of quality. I cannot take an absolutist point of view on this technology.

On the other hand, this technology is being instrumentalised from a nihilistic point of view to destroy so many vitally important cultural achievements. Both of these things are true, and acknowledging this does not make me a centrist.

I will always fight for technology that empowers people, disempowers the pointy-haired bosses of the world, and brings joy. What our students need is less hype and more 'hard engineering' of all kinds. I am endlessly optimistic about what we can do with the current cohort of students who are in some ways wiser than many of their teachers.

Josh Simmons Jan 26

@jonmsterling the only real point of difference i have there, is that i don't think the technology *can* exist outside the hype cycle. it only even vaguely functions when it's backed by truly colossal hardware resourcing, and only with the tremendous, constant human effort applied running & tuning it. i don't believe it can actually get much better in the current framework, and i don't think the small local models will be able to catch up either.

Josh Simmons Jan 26

@jonmsterling if that turns out to be wrong, i'll worry about it when it happens. :')

Jon Sterling Jan 26

@dotstdy That may be true. It's very possible that we are in the "golden age" (yuck) right now and it's all downhill from here. I'm not certain...

I am not technically up to speed with the state of the art in LLMs and competing models, etc. I am actually kind of optimistic that this stuff may become 10X cheaper in the future once the stranglehold of the LLM is broken, but I admittedly don't have a technically convincing argument for this. We'll have to see.

Joe Groff (1M Context)Jan 27

@jonmsterling @dotstdy it does seem to me that, especially within the confined search space of typed languages, we could get useful and predictable code generation with far fewer resources than LLMs need. though maybe that would require accepting a bit more hand-holding than typing in any vague natural language prompt and getting something back, or waiting for long periods of time without the illusion of feedback that chat provides

Colin Cornaby Jan 27

@joe @jonmsterling @dotstdy I've been pondering - but one issue is that you're usually writing an app that does *something.* If you want to vibe code a Chinese vocabulary tutor - the model needs to know Chinese vocabulary and teaching/lesson techniques.

If you toss vibe coding and go back to handholding like you suggest maybe possible. But that's a lot more handholding and it becomes much more necessary to get your hands dirty with code.

Joe Groff (1M Context)Jan 27

@colincornaby @jonmsterling @dotstdy Yeah that's the sort of thing about wannabe-omnipotent LLMs that's harder to beat, though (at least to me) also less interesting and more ambiguous how useful they really are. People seem to accept that they are getting reliable at mechanical code updates and tightly-scoped editing tasks, and the more specialized tool I can envision could do those sorts of things while still leaving the actual problem solving to the programmer who knows what they're doing

David Gerard Feb 8

@jonmsterling @dotstdy i am pretty sure that once the VC subsidy runs out, the APIs instead become 10x more expensive as they have to charge something they can make a profit on without subsidy

@davidgerard @jonmsterling @dotstdy you mean... it might become cheaper to employ meatbags that are also very competent at what they're doing?

HERESY!

Jeffy 🏳️‍⚧️🏳️‍🌈🇺🇦 ❤️🇱🇺Jan 27

@jonmsterling @mitchellh @nicklockwood

Like most analyses wanting to bothsides it, you fail to mention the water and pollution cost of it all.

@godofbiscuits @jonmsterling @mitchellh @nicklockwood

The resource issues will be partly (mostly?) solved by the new open source/weight models coming out from China; focusing on running on fewer resources &, as a side benefit, breaking the big US players' business models.

OTOH anyone using the current crop of LLMs (me included) are implicitly accepting that LLMs are allowed to pillage the knowledge and creativity of real people for the profit of a few oligarchs.

Konrad Hinsen Jan 27

@jonmsterling I share both your experience with young people and your view of generative AI.

The people most at risk, as far as I can tell, are young professionals who pick up AI tools for fun or efficiency without much reflection.

Leah Neukirchen @ Revision Jan 26

@jonmsterling The kids are alright.

Arcane Roops Feb 8

@jonmsterling @leah Good idea for a song :-)

CynthesisToday Jan 27

Thank you very much for sharing this perspective. I'm going to take this perspective, add it to the graphic from this article on stereographs to update my baseline worldview.

https://www.chartography.net/p/the-stereogram-strikes-back

@jonmsterling I remember when the java flavor of oo was crammed down everyone's throats in capitulation to industry interests, we didnt like that either but had no choice in the matter

Craig Brozefsky Feb 8

@deech @jonmsterling except that in this case it’s a religious ceremony of dedication to the brand of autocomplete you use to get wrong answers. 🤣

As a hiring manager, I am not going to expect you to be versed in fads that require thousands of dollars in tokens to learn that are of marginal value to producing good code, especially at the entry level. I also am applying my own criteria that you know how to find the documentation and perform experiments for the tools you use, and not listen to the bullshit bot.

Arcane Roops Feb 8

@jonmsterling They sound pretty well grounded--I mean they are there to learn a skill, and if they do use AI in future, that skill will help them to tell the wheat from the chaff.

Or the rest of us can look at
https://www.wheresyoured.at

Ed Zitron's Where's Your Ed At

The Words of Ed Zitron, a PR person and writer.

Ed Zitron's Where's Your Ed At

rainynight65 Feb 8

@jonmsterling I had a prof in Uni who fell under the 'clueless' category. He was supposed to teach Java, but everything he knew he got out of a book that he read over the summer break. Every time someone asked a question that went beyond the basics, his response was along the lines of "Great question. Why don't you read the documentation and tell us the answer next time?". Exasperating.

Everyday Freethought Feb 8

At my current employer, the More Bad Advice pinhead types all love it. The tech crowd is mixed. I think it will blow up either due to debt collapse or huge price increases that companies will be unwilling/unable to pay.

(That is before pointing out that anybody with at least two brain cells talking to each other should have rejected it the instant they understood what a hallucination is.)

Amongst the few non-tech people I know, most of them hate AI and wish it would go away.

@jonmsterling I've certainly heard mixed feelings on "AI" from students this year, compared to last year - and it's true that the most ardent believers that "AI is changing everything" are around my age [although it's not a universal belief in my cohort by any means, and there's at least as many strong AI skeptics there].

Robin Frost Feb 9

@jonmsterling Your post gives me hope for the future. #YayYoungPeople

🏳️‍⚧️PepperTheVixen🦯Feb 9

@jonmsterling Oh, it's definitely changing the world... for the worse. I 100% agree that we need to bring back real education and real jobs with real engineers

Maike Küper Mar 11

@jonmsterling oh good so I‘m just young at heart! Thanks for the relief :) I am so bored by the hype.